Influence of Maternal Denomination, God

Concepts, and Child-Rearing Practices

on Young Children’s God Concepts

SIMONE A. DE ROOS
JURJEN IEDEMA
SIEBREN MIEDEMA

This study tests a model of individual differences in God concepts among kindergarteners, based on social
learning and projection theory. Relations among maternal education, religious denomination, God concepts,
child-rearing practices, and young children’s God concepts were examined. Subjects were 363 Dutch preschoolers
(mean age = 66 months) and 271 of their mothers belonging to three religious denominations (open Christian,
orthodox Christian, and nonaffiliated). Child-rearing practices as well as God concepts were measured using
questionnaires. God concepts were operationalized as ideas about potential characteristics of God. The model
was partly supported. Maternal orthodox Christian denomination, God concepts, and child-rearing practices
all had effects on children’s “potent God” concept, confirming all parts of the model. Differences in children’s
“punishing God” concept were explained by strict child-rearing practices, providing evidence for projection
theory only. Children’s “loving and caring God” concepts were predicted by mothers’ “loving God” concept,
lending support for social learning theory.

INTRODUCTION

Although antecedents of God concepts have been a continuing source of fascination to psy-
chologists of religion, they have been little studied among children (cf. Heller 1986; Hertel and
Donahue 1995; Spilka, Addison, and Rosensohn 1975). Previous investigators of God concepts
among preschoolers and older children have, for the most part, concentrated on children’s age
differences, focusing mainly on the role of cognitive development in God concepts (cf. Fowler
1981; Nye and Carlson 1984; Smoliak 1999; Tamminen et al. 1988). However, we know rela-
tively little about the origins of individual differences in God concepts at a given age, especially in
young children. The study of the sources of the concept of God is important because it has often
been considered crucial in the development and formation of an individual’s personal faith (Hyde
1990). God concepts influence religious feelings and experiences, as well as devotional practices
(Hyde 1990; Tamminen 1991). God concepts may also indirectly become related to ethnic prej-
udice and political tolerance in adulthood via their possible association with aspects of religious
fundamentalism, theocratic beliefs, and religious commitment and participation (Karpov 2002;
Laythe et al. 2002). In addition, God concepts may have an effect on attitudes toward health. For
example, emphasis on God as a potent helper can positively affect children’s and adults’ coping
with chronic illness and death (Pendleton et al. 2002).
By God concept we mean subjects’ ideas concerning the different potential behavioral char-
acteristics of God, such as God as a loving, comforting, caring, potent, and/or punishing entity

Simone A. de Roos is Researcher in Educational Sciences, Faculty of Psychology and Education, Department of Philosophy
and History of Education, Free University Amsterdam, Van der Boechorststraat 1, 1081 BT Amsterdam, The Netherlands.
E-mail: SA.de.Roos@psy.vu.nl
Jurjen Iedema is Researcher and Methodologist in the Social and Cultural Planning Office, The Hague.
Siebren Miedema is Professor of Educational Foundations and Professor of Christian Education in the Department
of Philosophy and History of Education, and Dean of the Faculty of Psychology and Education, Free University
Amsterdam.

Journal for the Scientific Study of Religion 43:4 (2004) 519–535

520 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

FIGURE 1
A SOCIALIZATION THEORETICAL MODEL OF INTERRELATIONS AMONG
MOTHERS’ RELIGIOUS AND CHILD-REARING CHARACTERISTICS
AND CHILDREN’S GOD CONCEPTS

Maternal
Religious
Denomination

Maternal Children's
Childrearing God concepts

Maternal
God concepts

(cf. De Roos, Iedema, and Miedema 2001; Dickie et al. 1997). We refer to the Western Christian
tradition here.
Framed by a socialization theoretical model (see Figure 1), we propose that both social
learning (Bandura 1971; Beit-Hallahmi and Argyle 1997) and projection theory (cf. Rizzuto 1984;
Spiro and D’Andrade 1958) provide useful explanations for understanding individual differences
in young children’s God concepts. We will concentrate on the socialization of children by their
mothers, who, because they are often the primary caregivers, children have emotional ties with
and feel a natural dependence on (cf. Beit-Hallahmi and Argyle 1997). Socialization by mothers
was taken into account because women have been found to be more religious than men (Hyde
1990) and mothers probably have more impact on the development of their children’s religious
ideas than do fathers (Acock and Bengston 1978; Beit-Hallahmi and Argyle 1997; Tamminen
et al. 1988).
The model presupposes that mothers influence their children’s God concepts through so-
cial learning processes, here operationalized by a measure of maternal God concepts. Children’s
God concepts are learned, either deliberately or inadvertently, through the influence of parental
example (cf. Bandura 1971). In their religious instruction, mothers intentionally teach their chil-
dren about God. They provide models by explicitly talking about God and their views on God
to their children (cf. Herzbrun 1993). Children’s God concepts may also be formed through
more unintentional processes, such as symbolic and vicarious conditioning by mothers (Bandura
1971). Symbolic conditioning is the process by which children form evaluative reactions to-
ward God, certain things, or other people on the basis of little or no personal contact with
them. Such responses to God, an initial unknown, abstract, and invisible entity for children,
are probably developed by the connection of “God” with symbolic stimuli, often in the form of
emotion-arousing words used by parents. Children may develop positive or negative values of
God depending on the symbols with which God has been associated. For example, God may
be evaluated as pleasant by children through their mothers’ repetitive pairing of the word God
with adjectives having positive connotations, such as nice, kind, friendly. In vicarious learning,
the emotional responses of parents when talking about God are conveyed through vocal, facial,
and postural manifestations. This can arouse strong emotional reactions in the observing child,
reinforcing certain God concepts (cf. Bandura 1971). For example, mothers’ happy faces when
talking about a nurturing God may reinforce children’s idea of God as a trustful, loving, and caring
entity.
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 521

In the present study we use explicit maternal God concepts as an operationalization of social
learning and we expect young children to have God concepts that share the same features as
their mothers’ God concepts. In line with this expectation, among older children (from age
10 onward, see Tamminen 1991) and adolescents (Acock and Bengston 1978; Clark, Worthington,
and Danser 1988; Gibson, Francis, and Pearson 1990) similarities between parents’ and children’s
God concepts and religious beliefs, church attendance, attitudes toward Christianity, religious
experiences, and religious practices have been found. Only one study dealt with kindergarteners
(de Roos, Iedema, and Miedema 2001). This study showed that the more mothers perceive God
as a friend, the more their children will view God as a father or friend. A mother’s view of
God as distant was related to a less caring God concept among children. Children emphasized
God less as punishing than as loving (cf. Hertel and Donahue 1995). Furthermore, young children’s
punishing God was not associated with maternal God concepts. The subjects in this study were
mainly religiously nonaffiliated and “open” liberal Christians. Generally, mothers had low scores
on a strict view on God. The present research will examine whether the above-mentioned findings
will hold for a more diverse religious sample, including three conservative, orthodox Protestant
subdenominations. Since orthodox Protestants have been found to view God both as loving and
punishing (Stoffels 1995) and since a substantial number of our subjects are orthodox Christian, we
hypothesize a significant relation between maternal and children’s views of a strict and punishing
God.
The second way the model hypothesizes that mothers affect their children’s God concepts is
through their child-rearing practices (cf. Hyde 1990) (see Figure 1). Characteristics of maternal
child-rearing practices are thought to be projected by children onto their images of God (cf. Hertel
and Donahue 1995). However, findings concerning the association of child-rearing practices with
children’s and youth’s God concepts or other aspects of religion are inconclusive. Furthermore,
only one study has focused on preschoolers (Dickie et al. 1997). Some studies demonstrated
independent effects of child-rearing practices on aspects of children’s and youth’s religion (Hertel
and Donahue 1995; Nelsen 1981; Potvin 1977; Tamminen 1991). For example, parents’ anger
toward children (from age seven and older) was negatively related to children’s view of God’s
closeness, reality, care, and forgiveness (Tamminen 1991). Hertel and Donahue (1995) studied
both parental God concepts as well as perceived child-rearing attitudes as predictors of youth’s
(fifth through ninth graders) God concepts. They found that parents’ images of God as loving and
authoritarian, and youth’s images of parents as loving and authoritarian, predicted similar images
of God among the youngsters.
Other research has not found independent effects of child-rearing practices on children’s God
concepts and concepts of prayer (Dickie et al. 1997; Tamminen 1991; Worten and Dollinger 1986).
Dickie et al. (1997) studied, among other things, an aspect of child-rearing practices, specifically
parental discipline, as a predictor of children’s (4 to 11 year olds) God concepts. Children had to
rate their mothers’ and fathers’ discipline on love and power. Although love-oriented discipline
was positively and power-oriented discipline was negatively related to children’s perceptions of
God’s nurturance and power, type of discipline did not contribute independently to explaining
God’s power and God’s nurturance. Also, parental use of corporal punishment was not related to
a picture of God as stern, punishing, and frightening (Tamminen 1991).
The existence of the above-mentioned ambiguous findings may be due to differences in
methods used in assessing child-rearing practices, for example, parental reports (Tamminen 1991;
Worten and Dollinger 1986) versus children’s perceptions (Dickie et al. 1997; Hertel and Donahue
1995), in the operationalizations of child-rearing practices, such as parental discipline (Dickie
et al. 1997), child-rearing attitudes (Hertel and Donahue 1995), and corporal punishment (Nelsen
1981; Tamminen 1991), and/or to difference in the age of the children studied. It may be that
parental views on their child-rearing practices yield better results in the prediction of young
children’s God concepts than children’s perceptions of their parents’ methods of child rearing
(cf. Dickie et al. 1997). In the present study, we use maternal information about child-rearing
522 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

practices. We will focus on aspects of control and nurturance in the methods of child rearing and
expect strict and authoritarian child-rearing practices to be related to a punishing God concept
among the preschoolers. We expect that more loving and supportive child-rearing practices will
lead to a more loving perception of God among the children.
In our model we also assume an indirect link between the God concepts of mothers and
children, connected through child-rearing practices in which mothers may demonstrate (some of
the) characteristics they attribute to their God (cf. Hertel and Donahue 1995). If this hypothesis is
valid, there should be a connection between maternal God concepts and child-rearing practices.
Child-rearing goals and practices indeed have been found to differ as a function of religious be-
liefs. For example, it was found that parents who expressed the liberal religious belief that humans
are basically moral tend to be permissive in their child-rearing attitudes. Those who embraced the
belief that humans are at root immoral (a belief quite often shared with a literal interpretation of
the Bible and its teachings, which is common in conservative Christian circles) expressed au-
thoritarian child-rearing attitudes (Clayton 1988; Ellison and Sherkat 1993a). Parents’ view of
God as loving was associated with youth’s view of parents having loving child-rearing atti-
tudes (Hertel and Donahue 1995). Further support for the connection between religious beliefs
and child-rearing practices comes from cross-cultural research (cf. Lambert, Triandis, and Wolf
1959; Rohner 1975). These studies have shown that cultures in which loving, nurturing parent-
ing practices are predominant tend to have religious belief systems characterized by benevolent
deities, whereas cultures in which rejecting, authoritarian child-rearing practices predominate
tend to believe in more malevolent deities. In the present research we hypothesize that loving and
authoritarian views of God among mothers lead to nurturing and strict child-rearing practices,
respectively.
Finally, in predicting kindergarteners’ God concepts from maternal God concepts and child-
rearing practices, we control for maternal religious denomination since children’s God concepts
(de Roos, Iedema, and Miedema 2003), adults’ God concepts (Noffke and McFadden 2001), and
child-rearing practices (Alwin 1986; Danso, Hunsberger, and Pratt 1997; Ellison, Bartowski, and
Segal 1996; Ellison and Sherkat 1993a, 1993b; Kelley, Power, and Wimbush 1992) have been
found to differ as a function of denomination. In addition, we will control for (maternal) education,
which has been found to be related to child-rearing practices (Deković 1991; Rispens, Hermans,
and Meeus 1996) and aspects of religion (Alwin and Jackson 1982; Keysar and Kosmin 1995;
Lehrer 1999; Loury 2004).
We understand maternal denomination as an aspect of children’s socialization. Three main
religious denominations will be distinguished among the mothers: nonaffiliated; “open Chris-
tian,” consisting of Dutch Reformed and Catholic mothers; and “orthodox Christian,” containing
orthodox Reformed and Pentecostal mothers. Characteristic of the open Christian denomination
is the inclusive view that God’s revelation, as well as real experiences of God, are also found
in religions other than Christianity. However, the revelation through Jesus Christ is interpreted
as the ultimate means of salvation (cf. Miedema, 2000a, 2000b; Ziebertz 1994). The orthodox
Christian parents adhere to a strict, exclusive Christian faith, believing that their religious faith is
the only one that has the essential truth about humanity and deity (cf. Stoffels 1995). We expect
maternal denomination and God concepts to be interrelated (cf. Noffke and McFadden 2001)
without assuming a causal link. We expect that nonaffiliated mothers will have a more distant and
less potent and loving concept of God than affiliated mothers. Orthodox Christians are assumed
to have a more potent and punishing view on God than open Christians.
We also hypothesize that maternal denomination will affect child-rearing practices (cf. Danso,
Hunsberger, and Pratt 1997) and children’s God concepts (cf. de Roos, Iedema, and Miedema
2003). Orthodox Christian mothers are presumed to use more authoritarian child-rearing practices
than other mothers and we propose that they reinforce their children’s autonomy less than other
parents (Alwin 1986; Danso, Hunsberger, and Pratt 1997; Ellison, Bartkowski, and Segal 1996;
Ellison and Sherkat 1993a, 1993b; Wiley 1997). Children of orthodox Christian mothers are
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 523

expected to have a more punishing and potent view of God than do other children, whereas the
nonaffiliated children are hypothesized to have a less loving, caring, and potent God concept
than the other preschoolers (cf. de Roos, Iedema, and Miedema 2001, 2003). In summary, in this
article our goal is to test a model based on social learning and projection theory that predicts that
young children’s God concepts come from maternal God concepts and child-rearing practices,
controlling for education and religious denomination.

METHOD

Subjects

The subjects for this study were 363 Dutch preschoolers (mean age = 66 months) and 271 of
their mothers, sampled in two studies performed in 1999 and 2001 (92 mothers did not return the
questionnaires). In 1999, 165 kindergartners (mean age 64 months) participated along with 105 of
their parents (see de Roos, Iedema, and Miedema 2001). These respondents were distributed over
eight elementary schools belonging to four different religious denominations; 17 children of an
orthodox Reformed school in Amersfoort, 75 children of a Dutch Reformed school in Nijmegen
in the province of Zeeland, 26 children in two Catholic schools in Zeeland (southwest part of
the Netherlands), and 47 children in three state schools in Zeeland and Nijmegen. Amersfoort
and Nijmegen are medium-sized Dutch cities in the middle and east part of the Netherlands,
respectively.
In 2001, we examined 198 kindergarteners and 166 of their mothers. The mean age of these
children was 68 months. These children were recruited from six elementary schools belonging to
five different religious denominations and from one Pentecostal church. All children came from
the “Randstad,” a conglomoration of cities in the western part of the Netherlands. There were
49 children from two orthodox Reformed schools, 26 children from a Dutch Reformed school,
30 children from a Catholic school, 53 children in an interdenominational school (open Protestant
and Catholic), 20 children in a state school, and 20 children from a Pentecostal church. The
children from the Pentecostal church attended orthodox Reformed, Dutch Reformed, Catholic,
and state schools.
Among the mothers in both studies, 43.5 percent regarded themselves as churchless, 15.2 per-
cent were Dutch Reformed, 14 percent were Catholic (thus 29.2 percent were open Christian),
21 percent belonged to orthodox Reformed churches, and 6.3 percent belonged to a Pentecostal
church (thus 27.3 percent were orthodox Christian). The data collection for both studies took
place between March and July.

Measures and Procedures

Overview

Children’s and mothers’ God concepts were measured using interviews and questionnaires.
Maternal education, denomination, and child-rearing practices were studied employing a question-
naire. All children were interviewed individually by a female examiner in a 45-minute session
to assess their God concepts, as well as their concepts of self and of others and their attach-
ment representations (see de Roos, Miedema, and Iedema 2001). The children were taken from
their kindergarten classes to a separate room. The order of the different parts of the interview
session was the same for each child. First, in order to get acquainted with the interview pro-
cedure and experimenter, the children were asked to draw a picture of God and to tell some-
thing about the drawing. Second, they answered open questions about the nature of God (for
example, what is God, where is God, what is God able to do, etc.; see de Roos, Iedema, and
Miedema 2001). After that, the children completed scales referring to concepts of self and others.
524 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

Then the children were presented with a 23-item structured questionnaire concerning poten-
tial characteristics of God. Finally, the children’s mother-child attachment representations were
measured.

Scales for Child-Rearing Practices and Maternal and Children’s God Concepts

To reveal possible dimensions of maternal child-rearing practices and the God concepts of
both mothers and children, we conducted principal components explorative factor analyses with
varimax rotation on the items and responses for each measure. A minimum eigenvalue of 1.00 and
the scree test were used as criteria for extracting factors. Only items with absolute factor scores
>0.4 were selected. Based on each factor, two scales were constructed by averaging the subjects’
scores on each (randomly chosen) half of the constituent items. The two scales per factor were
used as indicator variables in a latent variable model (Arbucle and Wothke 1999). In this way,
we established how well the indicator variables represent the underlying latent variables. Using
this method the effect of measurement errors in the observed variables is eliminated (Jöreskog
and Sörbom 1993). The standardized coefficients of a latent variable to its indicator variables are
standardized validity coefficients (Bollen 1989), comparable to Cronbach’s α. These coefficients
are depicted in the last column of Table 1. Table 1 also shows the variables being used, with
numbers, means, ranges, and standard deviations.

Maternal Child-Rearing Practices

To assess maternal child-rearing practices, mothers completed a 29-item questionnaire that

is a Dutch version of the Childrearing Practices Report (CRPR) developed by Block (1965),
four items of a Dutch version of the Family Life in America (FLA) (Wolins 1963), and six
items of the Dutch questionnaire “Meningen Over de Opvoeding in het Algemeen”’ (MOA)
(General Beliefs about Childrearing) (Angenent 1974, 1976). Items were chosen on an empirical
(Deković, Janssens, and Gerris 1991; Rickel and Biasatti 1982; Siebenheller 1990; Trickett and
Susman 1988) and conceptual basis, with reference to maternal nurturance and control.
The FLA and MOA are questionnaires that use a six-point Likert-type scale. The Dutch
version of the CRPR used here consisted of 100 items (cf. Siebenheller 1990), an addition of nine
items to the original CRPR. The added items concerned parental restrictiveness and disciplinary
practices, for example: “My child has to obey me, no nonsense” (No. 97) or “When my child
behaves bad I will punish him/her by boxing his/her ear and slapping” (No. 95). Usually, the CRPR
is administered in a Q-sort format with a forced-choice distribution. Deković, Janssens, and Gerris
(1991) have shown that the CRPR can be used in a shorter questionnaire form simplifying the
administration, scoring, and interpretation of the CRPR (cf. Rickel and Biasatti 1982). For this
reason, the CRPR in the present study was employed just like the FLA and the MOA in the form
of a questionnaire using a six-point Likert-type scale, ranging from 1 (not at all descriptive of
me) to 6 (highly descriptive of me).
Factor analysis yielded four scales, which we called Strictness, Promoting Autonomy, Positive
Relationship, and Openness, and that explained 34.3 percent of the variance. Strictness refers to
child-rearing practices that are characterized by a high degree of parental control, narrow limits
on the child’s behavior, and authoritarian maternal discipline (11 items; CRPR item numbers 97,
1 (reversed loading), 70, 23, 14, 73, and 95; MOA item numbers 29, 36, 31, and 45). Promoting
Autonomy describes maternal encouragement of children’s independent activity, responsibility,
and problem-solving skills (nine items; CRPR item numbers 75, 6, 33, 41, 100, 67, and 21; MOA
item number 41; FLA item “I allow my child to make a lot of decisions”). Positive Relationship
shows the presence of affection, acceptance, and playful interactions, and the relative absence of
conflicts and disappointment in the mother-child relationship (five items; CRPR item numbers
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 525

TABLE 1
DESCRIPTIVE STATISTICS OF THE VARIABLES AND DIMENSIONS USED
(N = 271). IN THE LAST COLUMN: LOADINGS OF THE OBSERVED VARIABLES
ON THE LATENT VARIABLES (The Measurement Part of the Latent Variable Model)

Number of Cases

Vocational
Junior Secondary College or Standardized
Latent Variable Education Education University Coefficient

Mothers’ education 28 127 116 1∗

Maternal denominational affiliation∗∗
Open Christian (dummy) 79 1∗
Orthodox Christian (dummy) 74 1∗
Nonaffiliated (reference category) 118 –
Minimum Maximum Mean SD
Maternal loving God, scale 1 1 6 4.19 1.62 0.96
Maternal loving God, scale 2 1 6 4.31 1.68 0.99
Maternal strict God, scale 1 1 6 2.76 1.41 0.83
Maternal strict God, scale 2 1 6 2.51 1.28 0.91
Maternal distant God, scale 1 1 6 2.34 1.28 0.74
Maternal distant God, scale 2 1 6 2.14 1.59 0.75
Child rearing: strictness, scale 1 1.17 6 3.01 0.97 0.78
Child rearing: strictness, scale 2 1 5.2 2.79 0.86 0.84
Child rearing: autonomy, scale 1 1.4 5.6 3.48 0.74 0.74
Child rearing: autonomy, scale 2 1 5.75 3.42 0.75 0.65
Child rearing: posit. relationship, scale 1 2.67 6 5.18 0.64 0.71
Child rearing: posit. relationship, scale 2 1.5 6 4.39 1.08 0.61
Child rearing: openness, scale 1 3.33 6 5.32 0.51 0.76
Child rearing: openness, scale 2 1 6 5.43 0.61 0.55
Children’s potent God, scale 1 1 3 2.60 0.53 0.78
Children’s potent God, scale 2 1 3 2.68 0.50 0.66
Children’s punishing God, scale 1 1 3 1.80 0.80 0.83
Children’s punishing God, scale 2 1 3 1.81 0.96 0.81
Children’s caring God, scale 1 1 3 2.56 0.67 0.66
Children’s caring God, scale 2 1 3 2.51 0.54 0.84
Children’s loving God, scale 1 1 3 2.82 0.48 0.77
Children’s loving God, scale 2 1 3 2.64 0.60 0.85
Children’s God as like their parents, 1 3 1.42 0.79 0.58
1 item: God looks like daddy
Children’s God as like their parents, 1 3 1.26 0.66 0.74
1 item: God looks like mummy
∗
This variable has only one indicator variable with coefficient 1.
∗∗
Maternal denominational affiliation is a nominal variable represented by dummies for open Christian and
orthodox Christian mothers.

34, 40, 69 (reversed loading), 32 (reversed loading), and FLA item “My child and I often have
conflicts” (reversed loading)). Openness indicates the willingness of mothers to share both positive
and negative feelings and experiences with their children (five items; CRPR item numbers 58, 94,
42, 76, and 18).
526 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

Maternal Education, Denomination, and God Concepts

Mothers were asked their highest level of completed education (junior education, secondary
education, or vocational college and university) and their denominational affiliation (nonaffiliated,
open Christian (open Protestant and Catholic), and orthodox Christian (orthodox Protestant and
Pentecostal)). They completed a 25-item questionnaire about their own God concepts using a
six-point Likert scale. The items are intended to involve God’s nurturance and power (cf. Benson
and Spilka 1973), that is, “God is caring,” “God is comforting,” “God preserves the earth,” and
“God sees everything.”
We found three scales (explaining 70.8 percent of the variance), called Loving God (e.g., God
loves people, God is patient, God is caring, God loves me even when I do something against His
will, God is merciful; 15 items), Strict God (e.g., God condemns, God punishes, God is strict,
God controls me; seven items), and Distant God (e.g., God is aloof, God is not available, and God
is available for believing people only; three items).

Children’s God Concepts

To tap into the children’s God concepts, a 23-item questionnaire was used (cf. Dickie et al.
1997). The items are intended to involve different possible characteristics of God (i.e., God loves
me, God makes me happy, God sees everything you do, God is angry when you do something
bad). The experimenter read the items, and children rated each item on a three-point (no = 1,
sometimes = 2, or yes = 3) scale.
Five scales were distinguished (explaining 53.4 percent of the variance), called Potency of
God (e.g., God sees everything you do, God is the boss, God is very strong; seven items), Punishing
God (e.g., God punishes often, God punishes when you are naughty, and God is angry when you
do something bad; three items), Caring God (e.g., God helps people, God cares for people and
animals, God can comfort you when you’re sad; five items), Loving God (e.g., God loves me,
God makes me happy, God is a friend, and God is nice; four items), and God is like their parents
(e.g., God looks like daddy and God looks like mummy; two items).

Analysis Effects Study and Missing Mother Data

Before the research questions could be answered we first had to check whether subjects of
the 1999 study and the 2001 study differed with respect to maternal education, the answers on the
three dimensions of maternal God concepts, four dimensions of child-rearing practices, and five
dimensions of children’s God concepts. These dimensions (the two subscales of each measure
were summarized and averaged) were used as dependent variables in a multivariate analysis
of variance with study and maternal religious denomination (nonaffiliated, open Christian, and
orthodox Christian) as between-subjects factors. The main effect of study was not significant,
F(13, 215) = 1.28, ns, neither was the interaction effect between study and maternal religious
denomination, F(65, 1095) < 1, ns. Thus, the two studies seem comparable with regard to the
variables assessed here.
Ninety-two mothers did not return the questionnaires. They were excluded from the analyses,
leading to 271 children and their mothers without missing data. We compared children’s God
concepts between the group without mother data and the group with mother data in a multivariate
analysis of variance and found that there was no multivariate significant difference, F(5, 353) =
1.76, ns. To examine the research questions, structural equation modeling was used with the
program Amos (Arbucle and Wothke 1999).
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 527

RESULTS

The path diagram in Figure 2 shows the results of the structural equation modeling. The
model fits reasonably well: χ 2 (292) = 362.19, p = 0.003; normed fit index (NFI) = 0.98; relative
fit index (RFI) = 0.98; root mean square error of approximation (RMSEA) = 0.03 (satisfactory
values are a nonsignificant χ 2 , NFI > 0.90, and RFI > 0.90; RMSEA < 0.05 indicates a close
fit). The χ 2 is significant, but as Jöreskog and Sörbom (1993) have pointed out, it may not be
realistic to assume that the model holds exactly in the population and, therefore, in practice it
is more useful to regard χ 2 as a measure of fit rather than as a test statistic. In this regard, the
RMSEA is preferable as it takes account of the error of approximation in the population.
In Figure 2, latent variables are drawn in ellipses. Unobserved residual or error variables are
depicted in circles. Straight lines with one arrow represent an effect of one variable on another
(so-called paths). Curved lines with two arrows represent a covariation between two exogenous
(predictor) variables or an error covariation (a covariation between the residuals of two variables)
between two endogenous (dependent) variables. For maternal religious denomination, two latent
variables were used, comparing open Christians and orthodox Christians with the nonaffiliated as
the reference category. These latent variables have only one indicator variable and therefore the
measurement error cannot be estimated and is set to zero. The same holds for the latent variable
mothers’ education, also with only one indicator variable.
The final model shown in Figure 2 was created in a number of steps. First, we allowed the
exogeneous latent variables (the first column in Figure 2) to covary with each other (a common

FIGURE 2
STANDARDIZED ESTIMATES OF THE RELATIONS AMONG MATERNAL
EDUCATION, RELIGIOUS DENOMINATION, AND GOD CONCEPTS,
CHILD-REARING PRACTICES, AND CHILDREN’S GOD CONCEPTS
.29 -.14
Mothers' Childrearing: Children's
education Autonomy Potent God
-.23
e5
.13 -.28 e1
.27
.18
.21 .09
-.12 Open
Christian Mothers Children's .89
.26 Loving God
-.22 .18
.24 e6 .87
Childrearing:
Orthodox Strictness .83
Christian Mothers
Children's
e2 Caring God
.51 .25
-.16 e7
Maternal -.39
.37 .42
.63 Loving God e3 .19 .13
.54 .25
.59 Childrearing:
-.34 Children's e8
Positive Punishing God
Maternal Relationship
-.32
Strict God
-.48 .24
.19 e4
-.23
e9
Maternal Childrearing: Children's God
Distant God Openness like parents
528 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

routine for latent variable analyses). Next, child-rearing practices latent variables were allowed to
covary among each other, followed by the children’s God concepts latent variables. Nonsignificant
covariations were removed. Then, paths were drawn between the latent variables of, on the one
hand, maternal education and religious denomination and mothers’ God concepts, and, on the
other hand, child-rearing practices. Nonsignificant paths were removed. After that, paths were
traced between the latent variables of child-rearing practices and children’s God concepts. Again,
nonsignificant paths were removed. Next, paths were drawn between, on the one hand, the latent
variables of maternal education and religious denomination and mothers’ God concepts, and, on
the other hand, children’s God concepts, dropping the nonsignificant ones. Finally, modification
indices were used to examine whether significant and theoretically meaningful paths could be
added.
As can be seen in Figure 2 and in accordance with our expectations, differences were found in
mothers’ and children’s God concepts as a function of their religious denomination and education.
Open Christian mothers had a more loving and less distant view of God than did nonaffiliated
parents. Orthodox Christian mothers perceive God as more loving and more strict, and also as
less distant, than do the nonaffiliated ones. The orthodox mothers have children who see God as
more potent than other children. In addition, in line with the hypothesis, these mothers promote
children’s autonomy less than other mothers do, which in turn leads to a more potent God concept
among these children.
Mothers’ education was positively related to mothers’ view of a loving God and to an open
Christian denomination. It was negatively related to a distant perception of God. Mothers’ ed-
ucation was positively connected with the promotion of children’s autonomy, leading to a less
potent view on God among the kindergarteners. In addition, maternal education was negatively
associated with strict child-rearing practices, resulting in a less punishing and potent image of
God among the young children.
Our hypothesis that maternal God concepts contribute independently to children’s God con-
cepts was partly confirmed. As expected, mothers’ loving God concept was positively related
to children’s loving and caring God concept, as well as to children’s potent God concept. Fur-
thermore, mothers’ loving God led to the allowance of children’s autonomy, which in turn led
to a less potent perception of God among the children. However, children’s punishing God and
God is like their parents were not (directly) predicted by mothers’ God concepts. In line with our
hypotheses, a strict, authoritarian, and distant image of God among the mothers led to strict, au-
thoritarian child-rearing practices, resulting in more potent and punishing God concepts for their
children. The last finding and the relation found between autonomy and children’s potent God
concept confirms the expectation of independent effects of child-rearing practices on children’s
God concepts.
Mothers’ strict and distant God negatively predicted a positive mother-child relationship, but
this relationship had no effects on children’s God concepts.
A number of error correlations were found among child-rearing practices and preschoolers’
God concepts. Maternal strictness was negatively related to positive relationship and positively to
openness. Promoting autonomy was positively associated with openness. Children’s potent, lov-
ing, and caring God concepts were highly positively interrelated. Finally, small positive relations
were found among children’s punishing God concepts with children’s caring God and God is like
their parents concepts.

DISCUSSION

The present research tested a social learning and projection theoretical model of individual
differences in preschoolers’ God concepts by examining associations between maternal education,
denomination and God concepts, child-rearing practices, and young children’s God concepts.
Mothers’ and children’s God concepts were operationalized as ideas about potential characteristics
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 529

of God, such as God as a potent, punishing, caring, and/or loving entity. Mothers were either
nonaffiliated, open Christian (Catholic or Dutch Reformed), or orthodox Christian (conservative
Protestant). Maternal child-rearing practices were defined as mothers’ view on the strictness,
promotion of autonomy, positiveness, and the openness they use and experience in the relationship
with their child.
Employing causal modeling analysis, we found a significant latent variable model describing
many relations among the studied variables. We contribute to the parenting and religion literature
by providing a clearer understanding of the role of social learning and projection theory in the
formation of young children’s God concepts. Although the betas in the latent variable model
were often relatively small, they are statistically significant even after controlling for maternal
education and religious denomination.
Before we turn to the findings of the test of our model, we first will draw attention to the
predictability of the preschoolers’ different God concepts. Children’s caring and potent images
of God seem to be more strongly predicted by the antecedents studied here than were children’s
punishing God images. This finding is in line with results of previous studies (cf. de Roos,
Iedema, and Miedema 2001; de Roos, Miedema, and Iedema 2001; Hertel and Donahue 1995).
An explanation for this finding may be that a punishing God concept is less salient for young
Dutch children in the present time than a caring and potent God and, therefore, is less predictable
than other God concepts. This suggestion is congruent with results of empirical studies of the
last two decades that show an emphasis on the prevalence of God’s care over God’s authority
(cf. Hertel and Donahue 1995; Nelsen, Cheek, and Au 1985; Nelsen, Potvin, and Shields 1977).
The present study also shows that children score higher on a caring God (m = 2.53) and loving
God (m = 2.73) than on a punishing God image (m = 1.81). The suggestion is also conceptually
consistent with contemporary (post)modern theologies in which a loving God is emphasized and a
wrathful God is shifted away from (cf. Tieleman 1995; Tilley 1995). Moreover, it is consonant with
findings by Hutsebaut (1998) and Janssen and Prins (1998), who demonstrated that Belgian and
Dutch adolescents and adults construct a personal, changeable concept of God and do not express
images of a traditional, punishing, and wrathful God. So, our findings are, theoretically speaking,
fully in line with a constructionist concept of God as the result of a process of “bricolage.”
The individualization of the concept of God in the domain of private and social religion finds its
societal counterpart in the broad processes of individualization that have developed since the 1960s
(cf. Taylor 2002). It may be read as an example of metaphoric parallelism, that is, the association
between the supernatural (here, God) and society in which assertions about God are metaphoric
representations of social facts (cf. Durkheim’s [1915] 1964 perspective on religion as a social
construct; see also Hertel and Donahue 1995).
Children’s image of God as like their parents was not predicted by any variable in the present
study (cf. de Roos, Iedema, and Miedema 2001). Potential differences in children’s perceptions
as to whether God is like their parents probably are caused by age effects. In a recent study (de
Roos, Iedema, and Miedema 2003), for example, four year olds more often refer to their parents in
their idea of God than do five and six year olds. This was explained by the idea that the youngest
preschoolers are more dependent on their parents than the older ones (cf. Kirkpatrick 1999).
Now we attend to the test of our model. Findings concerning the prediction of children’s
potent God image confirmed all parts of our model. As anticipated by our hypothesis based on
social learning theory, maternal God concepts had an independent effect on young children’s
potent God concept. Mothers who have a loving idea of God were found to have children who
view God as potent. This result accords with previous research (de Roos, Iedema, and Miedema
2001; Hertel and Donahue 1995).
Following projection theory, strict, authoritarian child-rearing practices, and reinforcement
of autonomy, independently affected children’s potent image of God. Our results showed that the
more mothers reported strict child-rearing practices, the more their children have a potent concept
of God. Strict child-rearing practices encompass punitive as well as power-assertive characteristics
530 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

of parents that probably are attributed by the children to God. The less mothers stimulate their
child’s autonomy, the more their children tend to perceive God as potent. This finding may also
be interpreted as an example of social learning theory. It may be that children of these mothers are
taught that they can’t rely on themselves to learn to decide what is right and wrong, but that they
are dependent on a very powerful God and that they have to be submissive to Him (cf. Stoffels
1995). It seems that maternal information about child-rearing practices is more connected with
preschoolers’ God concepts than children’s own perception of parenting (cf. Dickie et al. 1997).
Concerning the effect of maternal denomination on young children’s potent God concept, we
found that children of orthodox Christian mothers have a more potent image of God than children
of nonaffiliated and open Christian mothers. A similar finding was reported by de Roos, Iedema,
and Miedema (2001, 2003) and is consonant with the hypothesis. However, contrary to the other
hypotheses, we did not find independent effects of mothers’ denomination on children’s loving,
caring, and punishing God concepts.
The results concerning associations of mothers’ denomination and God concepts with child-
rearing practices, and about the indirect link of maternal God concepts with children’s potent God
concept via child-rearing practices, are generally in line with our hypotheses and findings from
previous studies. Orthodox Christians were found to promote their children’s autonomy less than
other parents (cf. Alwin 1986; Danso, Hunsberger, and Pratt 1997; Ellison and Sherkat 1993a).
Orthodox mothers probably emphasize obedience and submission to a supernatural authority. In
line with this, children of these mothers probably are taught to obey and submit to the earthly
authority system (here, parents) that represents the supernatural authority. Orthodox mothers
seem to value obedience as a virtue, restricting the child’s autonomy (cf. Altemeyer 1996) and
resulting in a more potent view of God among the kindergarteners. Interestingly, we did not find
orthodox Christian mothers to have more strict, authoritarian child-rearing practices than other
mothers. This result is in accordance with recent studies suggesting that the disciplinary style of
conservative Protestant parents is not as authoritarian as has often been charged (Bartkowski and
Ellison 1995; Wilcox 1998). Although conservative Protestant caregivers are more likely than
other parents to rely on corporal punishment in child rearing, recent work reveals that they are
also more likely than other parents to praise and hug their children (Wilcox 1998) and are less
likely to report yelling at their children (Bartkowski and Wilcox 2000).
Mothers’ loving God concept was positively related to the promotion of their child’s auton-
omy, which in turn was negatively associated with children’s potent God concepts. This latter
result seems to contradict the finding that mothers’ loving God images independently lead to a
potent God concept among children. Thus, mothers’ loving God leads in a direct way to chil-
dren’s potent view of God, but it results in a less potent view via the simultaneous support of
children’s autonomy. Although these findings emerged across denominations, it may be that the
direct “positive” path of mothers’ loving God with children’s potent God is more applicable for
orthodox Christian families, whereas the indirect “negative” path may be more appropriate for
liberal Christian families. Liberal Christian families may teach their children that God loves you
as you are and that you are allowed to seek and find your own destiny, which may result in a less
potent image of God among the children.
Another explanation for these seemingly contradictory findings may be potentially different
connotations of “potency.” Although the items in the present potency scale mainly seem to refer
to a theistic, transcendent kind of potency (God as an almighty and omniscient majesty, as a
supreme ruler) it may be that for many children the items are connected with other connotations
for potency, for example, an immanent, pantheistic view, or a combination of transcendent and
immanent views, the so-called transcendence-in-immanence view (cf. Van der Ven and Biemans
1994). An immanent understanding of potency is, for example, (in grown-up language): “God
is an inner force in me that enables me to do the right things.” A mixture of both kinds of
interpretations of potency is, for example: “God has the power and is willing to feel our pain
and to cry with us and to give us unconditional love” and/or “God is the power who won’t
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 531

let the poor and underdogs be resigned to their fate” (cf. Caputo 2001; Schoonenberg 1991).
It may be that mothers’ image of a loving God is directly positively related to the latter two
interpretations of potency among children (immanent or transcendence-in-immanence), whereas
it may be negatively related to a transcendent form of potency via respect for children’s autonomy.
Further research is recommended to test these explanations more extensively.
Our results concerning the prediction of children’s loving and caring God concepts were in
line with only the social learning theoretical part of the model. The more mothers have a loving
concept of God, the more their children will perceive God as loving (including as a friend) and
caring. However, unlike our hypothesis and findings of de Roos, Iedema, and Miedema (2001),
a maternal distant God concept was not predictive of children’s caring God concept. Also, our
expectation based on projection theory was not confirmed, that is, we found no relations between
loving child-rearing practices and children’s loving and caring God concepts.
Findings about maternal influences on children’s punishing God concept are in accordance
with hypotheses based on projection theory only. Strict child-rearing practices lead to a punishing
image of God among the preschoolers. Hertel and Donahue (1995) reported similar findings
among older children. Remarkably, no associations were found between maternal and children’s
strict, punishing God images, whereas links between both generations for loving God concepts
were demonstrated. It may be that in the present time mothers express their views of a loving
and benevolent God more easily and straightforwardly to their children than a punishing God
concept; this may be a result of the appreciable decline in the popularity of images of God as
punitive (cf. Hertel and Donahue 1995; Nelsen, Cheek, and Au 1985; Nelsen, Potvin, and Shields
1977). Therefore, mothers may not explicitly talk about a punitive God, but show punishing God
concepts in their strict, coercive interactive behaviors with their children.
In addition, in accordance with our expectation about indirect links between maternal and
children’s God concepts via child-rearing practices, we found that those mothers who have a strict
image of God tend to be more inclined to mention strict child-rearing practices than other mothers,
resulting in a punishing image of God among the preschoolers. Hertel and Donahue (1995) and
Potvin (1977) reported comparable findings among older children and adolescents. We suppose
that mothers who perceive God as vindictive, stern, and wrathful emphasize children’s obedience
to God and to parents (e.g., Colossians 3:20), which is reflected in the endorsement of strict
child-rearing practices to enforce such obedience (cf. Danso, Hunsberger, and Pratt 1997). Also,
a distant God concept of mothers led to more strict child-rearing practices, which in turn led to
children’s punishing God concept.
Although we did find some connections between maternal God concepts and positive mother-
child relationships, this dimension of child-rearing practices did not predict any of the children’s
God concepts. This last finding is not coherent with findings of research among adults based
on attachment theory (cf. Kirkpatrick and Shaver 1992) or with Tamminen’s (1991) results that
showed that among 10 to 15 year olds a close relationship with parents was connected to the
concept of God as close, real, caring, and forgiving. As could be expected, the more mothers view
God as strict and distant, the less they report a positive mother-child relationship.
Mothers’ God concepts were not related to openness in maternal child-rearing practices.
Furthermore, mothers’ openness did not predict preschoolers’ God concepts. This may be due to
the relative low reliability of the openness scale (subscale 2: 0.55). That we found no connection
between positiveness and openness in the mother-child relationship with children’s God concepts
may also be caused by the social desirability of many items within these child-rearing dimensions.
Since the forced-choice format of the original child-rearing Q-sort was skipped, social desirable
responses among mothers may have been reinforced, which may obscure true associations among
the variables.
The last issues we would like to raise are the differences in maternal God concepts according
to religious denomination, and the relation of maternal education to aspects of maternal religion
and child-rearing practices. As hypothesized, orthodox Christian mothers had a more strict and
532 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

loving, and less distant, view of God than did nonaffiliated mothers. Open Christian mothers
describe God as more loving and less strict and less distant than nonaffiliated mothers. Although
we did not explicitly test this, on the basis of the covariances we may conclude that the open
Christian mothers have a less loving and less strict perception of God than the orthodox mothers.
These results are in line with those of Noffke and McFadden (2001). That nonaffiliated mothers
more often say God is not accessible or is only available for believers is understandable. A
remarkable and unexpected finding, however, is that the nonaffiliated mothers have a more strict
view of God than do the open Christian ones. An explanation may be that the nonaffiliated ones
doubt the existence of God due to the suffering in the world, don’t believe in God, or even blame
God for wars, disasters, and poverty (cf. de Roos, Iedema, and Miedema 2003). Therefore, they
view God as more punitive and less loving than do the open Christians. Further research is needed
to test this explanation.
With respect to the relation of mothers’ educational level with religious affiliation, God
concepts, and child-rearing practices, it was shown that the higher the educational level among
the mothers, the more these mothers belonged to an open Christian denomination, the more
they viewed God as loving, and the less they perceived God as distant. Comparable results were
reported by Lehrer (1999). Mothers with a higher educational level stressed the importance of
autonomy in their child rearing more, and reported less strict child-rearing practices, than those
having lower educational levels. The latter findings are in line with Rispens, Hermanns, and Meeus
(1996). The educational background of the majority of the mothers in the present study was a
secondary education, vocational study, or university. Additional research among mothers having
junior education is recommended to determine whether our findings are sample specific.
In conclusion, our social learning and projection theoretical model of individual differences in
young children’s God concepts was confirmed for children’s concepts of a potent God. Children’s
loving and caring God concepts were influenced by a maternal loving God concept, lending
support for the social learning theoretical part of the model. Finally, children’s punishing God
concept was predicted by strict child-rearing practices, providing evidence for our hypothesis
based on projection theory.

ACKNOWLEDGMENTS

The authors gratefully acknowledge Inez de Meester, Marloes van Delden-Verkerk, and Marieke van’t Zand for
taking part in the initiation of the research and the data collection.

REFERENCES

Acock, A. C. and V. L. Bengston. 1978. On the relative influence of mothers and fathers: A covariance analysis of political
and religious socialization. Journal of Marriage and the Family 40:519–30.
Altemeyer, B. 1996. The authoritarian spectre. Cambridge, MA: Harvard University Press.
Alwin, D. F. 1986. Religious and parental child-rearing orientations: Evolution of a Catholic-Protestant convergence.
American Journal of Sociology 92:412–40.
Alwin, D. F. and D. J. Jackson. 1982. The statistical analysis of Kohn’s measures of parental values. In Systems under
indirect observation part 1, edited by K. G. Jöreskog and H. Wold, pp. 197–223. Amsterdam, North Holland: Elsevier
Science Publishers.
Angenent, H. L. W. 1974. Opvoeding, persoonlijkheid en gezinsverhoudingen in verband met criminaliteit [Educa-
tion, personality, and family relationships associated with criminality]. Groningen: Criminologisch Instituut, Rijks
Universiteit Groningen.
———. 1976. De basisdimensies van de opvoeding, gemeten met een vragenlijst [Basic dimensions of childrearing
measured by a questionnaire]. Gedrag 4:39–51.
Arbucle, J. L. and W. Wothke. 1999. Amos 4.0 user’s guide. Chicago, IL: Small Waters.
Bandura, A. 1971. Social learning theory. Morristown, NJ: General Learning Press.
Bartkowski, J. P. and C. G. Ellison. 1995. Divergent models of childrearing in popular manuals: Conservative protestants
vs. the mainstream experts. Sociology of Religion: A Quarterly Review 56:21–34.
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 533

Bartkowski, J. P. and W. B. Wilcox. 2000. Conservative protestant child discipline: The case of parental yelling. Social
Forces 79:265–90.
Beit-Hallahmi, B. and M. Argyle. 1997. The psychology of religious behaviour, belief and experience. London/New York:
Routledge.
Benson, P. and B. Spilka. 1973. God image as a function of self-esteem and locus of control. Journal for the Scientific
Study of Religion 12:297–310.
Block, J. H. 1965. The child-rearing practices report. Berkeley, CA: Institute of Human Development, University of
California.
Bollen, K. A. 1989. Structural equations with latent variables. New York: Wiley.
Caputo, J. D. 2001. Religion. London/New York: Routledge.
Clark, C. A., E. L. Worthington, and D. B. Danser. 1988. The transmission of religious beliefs and practices to firstborn
early adolescent sons. Journal of Marriage and the Family 50:463–72.
Clayton, L. 1988. The impact of parental views of the nature of humankind upon childrearing attitudes. In The religion &
family connection: Social science perspectives, edited by D. Thomas, pp. 273–82. Prova, UT: Birmingham Young
University.
Danso, H., B. Hunsberger, and M. Pratt. 1997. The role of parental religious fundamentalism and right-wing authoritari-
anism in childrearing goals and practices. Journal for the Scientific Study of Religion 36:496–511.
Deković, M. 1991. The role of parents in the development of child’s peer acceptance. Unpublished doctoral thesis.
Nijmegen: Katholieke Universiteit Nijmegen.
Deković, M., J. M. A. M. Janssens, and J. R. M. Gerris. 1991. Factor structure and construct validity of the Block child
rearing practices report (CRPR). Psychological Assessment 3:182–87.
de Roos, S. A., J. Iedema, and S. Miedema. 2001. Young children’s descriptions of God: Influences of parents’ and
teachers’ God concepts and religious denomination of schools. Journal of Beliefs and Values 22:19–30.
———. 2003. Effects of mothers’ and schools’ religious denomination on preschool children’s God concepts. Journal of
Beliefs and Values 24:165–81.
de Roos, S. A., S. Miedema, and J. Iedema. 2001. Attachment, working models of self and others, and God concept in
kindergarten. Journal for the Scientific Study of Religion 40:607–18.
Dickie, J. R., A. K. Eschleman, D. M. Merasco, A. Shepard, M. Vander Wilt, and M. Johnson. 1997. Parent-child
relationships and children’s images of God. Journal for the Scientific Study of Religion 36:25–43.
Durkheim, E. [1915] 1964. The elementary forms of the religious life. London: George Allen & Unwin.
Ellison, C. G., J. P. Bartkowski, and M. L. Segal. 1996. Conservative Protestantism and the parental use of corporal
punishment. Social Forces 74:1003–28.
Ellison, C. G. and D. E. Sherkat. 1993a. Obedience and autonomy: Religion and parental values reconsidered. Journal
for the Scientific Study of Religion 32:313–29.
———. 1993b. Conservative Protestantism and support for corporal punishment. American Sociological Review 58:131–
44.
Fowler, J. W. 1981. Stages of faith: The psychology of human development and the quest for meaning. San Francisco,
CA: Harper & Row.
Gibson, H. M., L. J. Francis, and P. R. Pearson. 1990. The relationship between social class and attitude towards Christianity
among fourteen- and fifteen-year-old adolescents. Personality and Individual Differences 11:631–35.
Heller, D. 1986. The children’s God. Chicago, IL: University of Chicago Press.
Hertel, B. R. and M. J. Donahue. 1995. Parental influences on God images among children: Testing Durkheim’s metaphoric
parallelism. Journal for the Scientific Study of Religion 34:186–99.
Herzbrun, M. B. 1993. Father-adolescent religious consensus in the Jewish community: A preliminary report. Journal for
the Scientific Study of Religion 32:163–68.
Hutsebaut, D. 1998. Omgaan met geloof en godsdienst: Ontwikkeling of bepaald door opvoedingsmodel? Een analyse
van gegevens van personen tussen 18 en 90 jaar [Dealing with faith and religion: Development or affected by a
model of childrearing? An analysis of data from subjects aged between 18 and 90 years]. In Schering en inslag.
Opstellen over religie in de hedendaagse cultuur, edited by J. Janssen, R. van Uden, and H. van der Ven, pp. 18–31.
Nijmegen: KSGV.
Hyde, K. E. 1990. Religion in childhood and adolescence: A comprehensive review of the research. Birmingham, AL:
Religious Education Press.
Janssen, J. and M. Prins. 1998. “Let’s reinvent the Gods.” De religie van Nederlandse jongeren in een Europese context
[The religion of Dutch adolescents in an European context]. In Schering en inslag. Opstellen over religie in de
hedendaagse cultuur, edited by J. Janssen, R. van Uden, and H. van der Ven, pp. 123–38. Nijmegen: KSGV.
Jöreskog, K. G. and D. Sörbom. 1993. Lisrel 8: Structural equation modelling with the SIMPLIS command language.
Hilldale, NJ: Lawrence Erlbaum Associates.
Karpov, V. 2002. Religiosity and tolerance in the United States and Poland. Journal for the Scientific Study of Religion
41:267–88.
Kelley, M., T. Power, and D. Wimbush. 1992. Determinants of disciplinary practices in low-income black mothers. Child
Development 63:573–82.
534 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION

Keysar, A. and B. A. Kosmin. 1995. The impact of religious identification on differences in educational attainment among
American women in 1990. Journal for the Scientific Study of Religion 34:49–62.
Kirkpatrick, L. A. 1999. Attachment and religious representations and behavior. In Handbook of attachment: Theory,
research, and clinical applications, edited by J. Cassidy and P. R. Shaver, pp. 803–22. London/New York: Guilford
Press.
Kirkpatrick, L. A. and P. R. Shaver. 1992. An attachment-theoretical approach to romantic love and religious belief.
Personality and Social Psychology Bulletin 18: 266–75.
Lambert, W. W., L. M. Triandis, and M. Wolf. 1959. Some correlates of beliefs in the malevolence and benevolence of
supernatural beings: A cross-societal study. Journal of Abnormal and Social Psychology 58:162–69.
Laythe, B., D. G. Finkel, R. G. Bringle, and L. A. Kirkpatrick. 2002. Religious fundamentalism as a predictor of prejudice:
A two-component model. Journal for the Scientific Study of Religion 41:623–35.
Lehrer, E. L. 1999. Religion as a determinant of educational attainment: An economic perspective. Social Science Research
28:358–79.
Loury, L. D. 2004. Does church attendance really increase schooling? Journal for the Scientific Study of Religion 43:119–
27.
Miedema, S. 2000a. The aims for religiously inspired urban schools. Dialogue, solidarity, and personal identity formation.
International Journal of Education and Religion 1:89–99.
———. 2000b. The need for multi-religious schools. Religious Education 95:285–98.
Nelsen, H. M. 1981. Gender differences in the effects of parental discord on pre-adolescent religiousness. Journal for the
Scientific Study of Religion 20:351–60.
Nelsen, H. M., N. H. Cheek, and P. Au. 1985. Gender differences in the images of God. Journal for the Scientific Study
of Religion 24:396–402.
Nelsen, H. M., R. H. Potvin, and J. Shields. 1977. The religion of children. Washington, DC: U.S. Catholic Conference.
Noffke, J. L. and S. H. McFadden. 2001. Denominational and age comparisons of God concepts. Journal for the Scientific
Study of Religion 40:747–56.
Nye, W. C. and J. S. Carlson. 1984. The development of the concept of God in children. Journal of Genetic Psychology
145:137–42.
Pendleton, S. M., K. S. Cavalli, K. I. Pargament, and S. Z. Nasr. 2002. Religious/spiritual coping in childhood cystic
fibrosis: A qualitative study. Pediatrics 109(1):1–11.
Potvin, R. H. 1977. Adolescent God images. Review of Religious Research 19:43–53.
Rickel, A. U. and L. L. Biasatti. 1982. Modification of the Block child rearing practices report. Journal of Clinical
Psychology 38:129–34.
Rispens, J., J. M. A. Hermanns, and W. H. J. Meeus, editors. 1996. Opvoeden in Nederland [Childrearing in the
Netherlands]. Assen: Van Gorcum.
Rizzuto, A.-M. 1984. The birth of the living God. Chicago, IL: University of Chicago Press.
Rohner, R. P. 1975. They love me, they love me not. New Haven, CT: HRAF Press.
Schoonenberg, P. 1991. De Geest, Het Woord en de Zoon. Theologische overdenkingen over Geest-christologie, logos-
christologie en drieëenheidsleer [The spirit, the word, and the son. Theological reflections on spirit-christology,
logos-christology, and the doctrine of trinity]. Kampen: Kok/Averbode, Altiora.
Siebenheller, F. A. 1990. Problematische opvoedingssituaties: Percepties, emoties en disciplineringsreacties van ouders:
Een interactionistische exploratie van antecedenten van reacties van ouders in alledaagse problematische opvoed-
ingssituaties [Problematic childrearing situations: Parental perceptions, emotions, and disciplinary reactions: An
interactional exploration of antecedents of parental responses in everyday problematic childrearing situations].
Zutphen: Thieme.
Smoliak, W. G. 1999. What children’s narrative tells us about their developing thoughts about God. In Being human:
The case of religion—Volume 2. Spiritual and religious development, edited by K. H. Reich, F. K. Oser, and W. G.
Scarlett, pp. 83–100. Lengerich, Germany: Pabst Scientific Publishers.
Spilka, B. J., J. Addison, and M. Rosensohn. 1975. Parents, self, and God: A test of competing theories of individual-
religion relationships. Review of Religious Research 16:154–65.
Spiro, M. E. and R. G. D’Andrade. 1958. A cross-cultural study of some super-natural beliefs. American Anthropologists
60:456–66.
Stoffels, H. 1995. Als een briesende leeuw: Orthodox-protestanten in de slag met de tijdgeest [Like a roaring lion:
Orthodox Reformed people dealing with the spirit of the age]. Kampen: Uitgeverij Kok.
Tamminen, K. 1991. Religious development in childhood and youth: An empirical study. Helsinki: Suomalainen
Tiedeakatemia.
Tamminen, K., R. Vianello, J. M. Jaspard, and D. Ratcliff. 1988. The religious life of preschoolers. In Handbook of
preschool religious education, edited by D. Ratcliff. Birmingham, AL: Religious Education Press.
Taylor, C. 2002. Varieties of religion today. William James revisited. Cambridge, MA and London, UK: Harvard University
Press.
Tieleman, D. 1995. Geloofscrisis als gezichtsbedrog. Spiritualiteit en pastoraat in een postmoderne cultuur [Crisis in
faith as an optical illusion. Spirituality and pastoral care in a postmodern culture]. Kampen: Uitgeverij Kok.
INFLUENCES ON YOUNG CHILDREN’S GOD CONCEPTS 535

Tilley, T. W., editor. 1995. Postmodern theologies. The challenge of religious diversity. Maryknoll/New York: Orbis
Books.
Trickett, P. K. and E. J. Susman. 1988. Parental perceptions of child-rearing practices in physically abusive and nonabusive
families. Developmental Psychology 24:270–76.
Van der Ven, J. A. and B. Biemans. 1994. Religie in fragmenten: Een onderzoek onder studenten [Religion in parts: A
study among students]. Kampen: Kok and Weinheim: Deutscher Studien Verlag.
Wilcox, W. B. 1998. Conservative protestant childrearing: Authoritarian or authoritative? American Sociological Review
63:796–809.
Wiley, A. 1997. Religious affiliation as a source of variation in childrearing values and parental regulation of young
children. Mind, Culture, and Activity 4:86–107.
Wolins, M. 1963. Selecting foster parents: The ideal and the reality. New York: Columbia University.
Worten, S. A. and S. J. Dollinger. 1986. Mothers’ intrinsic religious motivation, disciplinary preferences, and children’s
conceptions of prayer. Psychological Reports 58:218.
Ziebertz, H. G. 1994. Religionspadagogik als empirische Wissenschaft [Religious education from an empirical point of
view]. Weinheim: Deutscher Studien Verlag.